Wednesday, January 13, 2010

General Nature of Economic Model Dependency Schema

Following on the heels of my previous post, this essay will classify the different types of economic predictions and trace them back to their source data by following the paths that this data takes through the models (dependency scheme). In keeping with this approach, the objective of this study is to evaluate the degree to which various assumptions play roles in the actual conclusions observed within a given model. Once the structure of the model is understood it can be critiqued as being more or less scientific on the basis of the criteria described in my previous post: correspondence between model elements and observed reality, the degree to which it addresses the original questions of its scope, and the degree to which it behaves as a way of observing through the creation of structure of reality theories.

Below is a schematic depicting the dependencies for a classical economic model.



Three important observations can be made:
  1. Wages do not alter commodity demand in this model, as decreases in wages mean increases in profits (not separated schematically), which in turn increases in other factors. Similarly for wage increases. This is an example of the type of troublesome details that seem unavoidable in models built this way.
  2. This model, though predicated on the use of calculus as maximization, does not feature equilibrium in the sense of related rates. Thus, the model is "always in equilibrium", with equilibrium being nothing more than the (assumed unique and extant) solution to a linear system of constraints.
  3. Exogenous factors disturb a key point in the system - commodity demand, which feeds both the aggregate supply and the aggregate demand side of the equilibrium. Unless some means is devised ad hoc for bounding or predicting these exogenous factors, the predictive power of the model is compromised.
Generally, similar criticisms can be made of other economic models.

The key scope question asked by macroeconomics is How do changes in Government policy affect the parts of the economy? This question, in turn, spawns the following questions: How are the parts of the economy best described in terms of their functional relationships? And, What effects are economically relevant when reviewing or proposing policy changes? In addition, the first question can be pushed back to encompass brainstorming by asking What changes are occuring in the economy under the status quo? And What policies should be implemented to address changes in the economy?

Rather than addressing the question of whether a model matches observations as a whole, it is important to ask whether this model matches the behavior of its component groups. Generally, one should ask how the component groups behave when analyzed in this model, then examine the behavior of individuals and firms based on historical records and experimental data. Having fewer variables involved reduces the effect of exogenous factors. Additionally, without such analytic separation, progress is rendered impossible when particular errors cannot be identified because the solution space is too large to practically check them all.

Next a system of accumulation and exhaustion rates must be formulated to allow the model to have equilibrium generated not by constraint but by the equality of countervailing rates. This introduces the temporal element to the equations, and allows the direct description of economic trends, growth, etc. When these things are directly described, they may be directly compared to observations.

Finally, the development of metrics must go beyond mere measures of aggregate production. Single index values are not acceptable because many ethical systems are predicated on deontological standards such as minimums, rights, duties, and social orders. These must be integrated into the economic variables in such a way as to preserve their original meaning, and this requires knowledge of both the relevant philosophy and the economic model. It is important that this development be scientific, because in this way it can remain based on observations, allowing critique not merely of the economic methods but the philosophic ones as well.

No comments: